Dr. Ji Ping Shi Air Quality Modelling and Assessment Unit

Environment Agency

Modelling of ammonia dispersion and deposition from intensive farming

ADMLC meeting 9 February 2011

Ji Ping Shi and Colin Powlesland

Questions

� Why do we need this technical guidance?

� What is the technical thinking behind the guidance ?

Structure of presentation

� Why do we need to model ammonia from intensive farming

� What are the problems in the modelling application

� How have we dealt with the problem� Application of the recommended method

Legislation

� Applications for authorisation under the PPC / EPR regulations will need to be assessed in relation to the requirements of the Habitats Directive and Birds Directive and implementing UK regulations (The Conservation of Habitats and Species Regulations 2010)

Environmental Permitting Regime� To determine applications for permits under

the Environmental Permitting Regime information is needed on the potential environmental impact of the installation being considered.

Impacts of Ammonia� Direct damage to sensitive species

(e.g. leaf discoloration)� Changes in species composition - reduced

species richness� Increased sensitivity to other stresses (e.g.

frost)� Acidification of soils and freshwaters� Other effects (e.g. slime)

Damage to lower plants

Healthy

Damaged by ammonia

Lichen Moss

Problems encountered

Comparison without deposition

Comparison without deposition

AERMOD and ADMS deposition modules

� AERMOD and ADMS have different deposition algorithms, especially in the surface resistance

� But are there any other reasons causing the significant discrepancy in the prediction as shown above? � For example, ammonia deposition velocity?

Review Ammonia deposition velocity� Literature review shows that the reported

ammonia dry deposition velocity varied from 0.13 to 4.16 cm/s for grassland and moorland/heathland, from 0.79 to 6.56 cm/s for forest.

� J N Cape et. al., presented a paper at Edinburgh UNECE Ammonia Workshop (2006) which set the new critical levels.

� The paper suggested 1.6-3.2 cm/s for short vegetation and 3.3 – 4.8 cm/s for tall vegetation. These ranges assumed based on the annual average wind speed dependent data for the UK for each 5 km grid square.

Variation of Vd with ammonia concentration

� The majority of the above studies were carried out in the ambient ammonia concentration.

� However, when close to a pig/poultry farm, the ammonia concentration, annual mean and hourly mean can be very high.

� Studies show that ammonia deposition velocity decreases with increase of ammonia concentration

Deposition velocity vs long- term ammonia concentration

An empirical approachRecommend ammonia dry deposition velocities at

different long term average concentrations to be used in an impact assessment.

NH3 conc

(µg/m3)< 10 10 -

20 20 -

30 30 –

80> 80

Deposition velocity (m/s)

0.02 - 0.03*

0.015 0.01 0.005 0.003

* 0.02 m/s for short vegetation, and 0.03 m/s for tall vegetation.

Neil Cape, from CEH, suggested a similar approach, with deposition (g/m2/d) expressed as:

� <0.02 for concentrations < 10 ug/m3 [scaled in proportion to concentration]

� 0.02 +/- 0.01 for 10-80 ug/m3

� >0.02 for concentrations > 80 ug/m3 [scaled in proportion to concentration]

Vd vs Conc

0

0.01

0.02

0.03

0.04

0 10 20 30 40 50 60 70 80 90 100Ammonia conc (ug/m3)

Vd (m

/s)

EANeil CapeUpper rangeLower range

Uncertainty� The recommended approach is based on studies into

long-term observation of the relation between ammonia deposition and concentration. The use of the long term average concentration avoids the complexity encountered in short term (i.e., hourly) modelling.

� The recommended deposition velocities are based on the limited data available. There is currently insufficient information to do an uncertainty analysis.

� However, the proposed empirical approach does allow both the concentration dependency of the deposition velocity and the ammonia depletion of the plume to be taken into account.

Application - Deposition modelling with variable deposition velocity

� AERMOD ?

� ADMS ?

� Fd = Vdg C� Vdg = (Ra + Rb + Rc )-1

A tm osphe ric Source

A e rodynam ic, R a

Sublaye r, R b

Ove

rall

Sur

face

, Rc

Gro und, R g

Stom atal, R s /LA Ir

Cuticular, R cut/LA Ir

Mesophyll, R m /LA Ir

A e rodynam ic to Ground, R ac

So il, W ate r, e tc.

Scheme of major components of resistances used for gaseous dry deposition

Rc = [LAIr(Rs + Rm)-1 + LAIrRcut-1 + (Rac + Rg) -1] -1

,

Where LAIr = relative leaf area index (unitless).

AERMOD Gas Deposition

ADMS deposition module

� User defines a Vdg , Fd = Vdg C

� When Vd is unknown,r = ra + rb + rs

Type Definition Examples Rc

(s m-

1)Reactive

gasesGases expected to

undergo significant chemical

reaction with the surface

SO2

, O3

, NOx

, I2

, Cl2

, HF

30

Unreactive gases

Gases not undergoing significant chemical

reaction with the surface

CO2

, CH3

I 1000

Inert gases Noble gases He, Ne, Ar, Kr,

Xe,Rn

∞

Modelling assessment with deposition velocity varying with concentration

� ADMS 4.2 -- Spatially varying deposition

� This is not an ideal way for the calculation of deposition varying with concentration, but it makes the assessment possible.

� Question to model developers

-Single point source (1 g/s) - H = 3m, exit velocity = 0.001m/s - Met data: Marham 2007

-

Point source (1 g/s)-

H = 3m, exit velocity = 8 m/s

-

Met data: Marham

2007

Ji Ping Shi and Colin Powlesland

Acknowledgement

We would like to thank to following people for their useful discussion and comments during the guidance development:

Malcolm Sharp from SAC Consulting

Steve Smith, Ian Barrie from ADAS

Neil Cape, Ron Smith, Bill Bealey and Mark Theobald from CEH